Receipt Image Analysis System for Variable Format Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information analysis systems face inefficiencies when processing multiple receipts of varying shapes and sizes, requiring complex operations from users, and lack functionality for easy electronic storage and analysis.
Innovation Solution
An information analysis system comprising a reading device, an analyzer, and a display device, which includes a scanner, transmitter, processor, and memory to scan, transmit, and analyze receipt images, perform character recognition, and display results, allowing for efficient processing and storage of multiple receipts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a fixed shape reading method is used for receipts, then the processing method is simple, but it cannot handle multiple receipts with different shapes and sizes efficiently
Solution Approach 1:
The system segments the receipt processing into independent units by detecting and extracting individual receipt images from the scanned document. Each receipt is processed separately through character recognition and information extraction, allowing the system to handle varying shapes and sizes without requiring complex manual operations. The segmentation is achieved through image processing that identifies receipt boundaries and extracts them as separate entities for individual analysis.
2Measurement precision
If character recognition is performed on each extracted receipt image, then information extraction accuracy is improved, but processing time increases
Solution Approach 1:
The system performs preliminary actions by pre-processing each extracted receipt image before character recognition. This includes adjusting image quality, correcting skew angles, and optimizing image parameters to ensure accurate character recognition. By preparing the images in advance and using optimized recognition parameters, the system achieves high accuracy while minimizing the time required for the actual recognition process.
3Productivity
If the system processes receipts sequentially, then the processing logic is simple, but the overall processing efficiency is reduced
Solution Approach 1:
The system merges multiple receipt processing operations into a unified automated workflow. Instead of handling receipts sequentially with complex manual controls, the system combines image scanning, receipt extraction, character recognition, and information extraction into a single integrated process. This merging of operations enables parallel processing of multiple receipts, significantly improving productivity while the automated control logic manages the complexity of coordinating these operations.
Data Source
AI summary
A reading device scans receipts, and transmits image data of the receipts to an analyzer. The analyzer receives the image data, extracts images of the receipts from the image data; counts a number of the images; executes a character recognition process to obtain text data; transmits, to the display device, a number of all of the images, and sequentially transmits, to the display device, the images for which the character recognition process has been completed, together with the text data. The display device sequentially displays each image, together with the text data; and upon detecting that the number of all of the images matches a number of the sequentially transmitted images, receives a command to report to the analyzer that the reception the images is completed.